Human retroviral sequences associated with extracellular particles in autoimmune diseases: epiphenomenon or possible role in aetiopathogenesis?

Microbes and Infection, 1, 1999, 309−322 © Elsevier, Paris

Review

Human retroviral sequences associated with extracellular particles in autoimmune diseases: epiphenomenon or possible role in aetiopathogenesis? H. Perrona,b*, J.M. Seigneurinb a BioMérieux SA, Chemin de l’Orme 69280 Marcy l’Etoile , France Laboratoire de virologie médicale moléculaire, Faculté de Médecine de Grenoble, Domaine de la merci, 38700 La Tronche, France

b

ABSTRACT – Publications describing retroviral sequences associated with extracellular particles in Sjögren’s syndrome or systemic lupus erythematosus, multiple sclerosis, and type I diabetes present novel arguments and raise complex questions about eventual relationships between retroviruses and autoimmunity. They are presented and discussed in the present review, preceded by an overview of the biology of retroviral elements. © Elsevier, Paris retrovirus / superantigen / autoimmunity / multiple sclerosis / diabetes / systemic lupus

1. Introduction Several recent publications have reported the identification of retroviral sequences associated with virions produced by cells of patients with ‘autoimmune’ diseases: HRV-5 in Sjögren’s syndrome (SS) [1], IDDMK in type 1 diabetes or insulin-dependent diabetes mellitus (IDDM) [2], and MSRV in multiple sclerosis (MS) [3]. Over the past few years, numerous review articles have argued in favour of a potential role of retroviruses, particularly endogenous retroviruses (ERVs), in the aetiopathogenesis of autoimmune diseases [4]. In this review, we will discuss recent data concerning the search for retroviral sequences associated with retrovirus-like particles previously described in several human autoimmune diseases. However, since the characterization of these retroviral genomes also raises the question of endogenous retroviral elements – ERVs – and in order to better define the concepts in which ERVs are involved, we have summarized basic data concerning ERV origin and genetics in the first part of this review.

2. Endogenous retroviruses: genes or viruses? 2.1. Classifications of retroviruses

Retroviruses can be classified according to different criteria. Usually, they are classified into three subfamilies: * Correspondence and reprints Microbes and Infection 1999, 309-322

the lentiviruses or Lentivirinae (e.g., human immunodeficiency viruses, HIVs; visna retrovirus of sheep, causing chronic demyelinating disease), the oncoviruses or Oncovirinae (e.g., human T-lymphotropic retroviruses, HTLV, causing leukaemia or chronic demyelinating disease; murine leukaemia viruses, MLVs; mouse mammary tumour viruses, MMTVs) and the spumaviruses or Spumavirinae (e.g., simian foamy virus, apparently nonpathogenic but causing cytopathic effect in vitro, characterized by ‘foamy’ cells). They can also be classified into two categories: simple and complex retroviruses, depending on the existence of regulation genes expressed by multiple splicing of retroviral RNAs usually reassembling fragments within the 3’ part of the genome (figure 1). Another classification can differentiate between two other categories: exogenous retroviruses, which are ‘classical’ viruses transmitted by infection (horizontally by contagion or contamination, or vertically by in utero or perinatal infection) and endogenous retroviruses or ‘retroviral elements’, which normally have a mendelian ‘hereditary’ transmission (via chromosomes in gametes), but may also comprise replication-competent copies generating virions with potential infectivity and horizontal transmission. It is noteworthy that exogenous retroviruses classified in the ‘complex’ category have rather limited sequence homology with inherited endogenous sequences within the host’s species (e.g., HTLV-1 with human retroviral endogenous sequence 1 in part of gag only; cf. table I) 309

Review

Perron and Seigneurin

Figure 1. Possible phylogenic origin of retrotransposable elements and retroviruses. (adapted from H. Temin [6]). pol, polymerase gene encoding reverse transcriptase; LTR, long terminal repeat; gag, gene encoding nucleocapsid proteins; env, gene encoding envelope glycoproteins; regs, regulation genes expressed after muliple splicing of coding fragments in the 3’ part of the genomic RNA.

making the provirus detection in cellular DNA quite unaffected by nonspecific detection of homologous cellular sequences. Conversely, quite all exogenous retroviruses classified in the ‘simple’ category have strong and extended homologies over their entire genome with genetically related families of endogenous retroviral sequences (e.g., MLV and MMTV). In addition, they all belong to the oncovirus subfamily. Consequently, most endogenous retroviruses (ERVs) are genetically related to the oncoviruses. ERVs are present in the normal genome of all animal species. Their presence could be explained by a succession of phenomena which will be dicussed below. 2.2. Genetically transposable elements

Retroviruses belong to the genetically ‘transposable’ elements. These elements also comprise less complex genetic structures, such as the retrotransposons, which do not contain any ‘env’ gene, or such as the retroposons, which do not contain long terminal repeats either [5, 6]. The main characteristic of the retrotransposable genetic elements consists of the presence of a reverse transcriptase gene. A proposed ‘endocellular’ phylogenic origin of 310

these elements is presented in figure 1. Taken altogether, retroelements represent about 10% of the human genome, and ERVs per se, about 1% [5]. More rudimentary elements, classified in ‘transposable’ elements, also exist which do not comprise any reverse transcriptase gene (e.g., short interspersed repetitive elements, Sines) but have flanking repeats. 2.3. Evolution of retrotransposable elements

According to H. Temin [6], the evolution of retrotransposable genetic elements towards retroviruses could begin with an ancestral reverse transcriptase (RT) encoded by a polymerase (pol) gene, followed by the acquisition of a a ‘core protein’ gene (gag, for ‘group antigen’) with a flanking repeat (long terminal repeat), possibly through a transposable element, and of a glycoprotein gene (env, for ‘envelope’), as illustrated in figure 1. According to certain authors, the origin of the RT gene could be traced as far back as the events coinciding with the procaryotic invasion of primitive eucaryotes [7]. In parallel with retrotransposable elements, genes encoding ‘regular’ cellular enzymes such as telomerase are homologous to an RT Microbes and Infection 1999, 309-322

Retroviral particles and sequences in autoimmune diseases

Review

Table I. Main human endogenous retroviral families. Retrovirus 4–1 51–1 ERV-1 ERV-3 HLM-2 HERV-K HuRRS-P RTVL-H HRES-1 M7-rel S71 ERV-9 Hs5 EHS-1 EHS-2

Length (kb)

Number of copies

8.8 6 3–4 9.9 9 9.5 8–9 5.8 5–6 4–10 6 8 5–10 3.2–9.5 3.2–9.5

5–100 25–50 1 1 50 50 20–40 1000 1 300 1 35–40 ? ? ?

Homology with exogenous retroviruses MoMuLV MoMuLV MoMuLV/BaEV BaEV MMTV MMTV MoMuLV ? HTLV-1 M1 baboon RV SSAV ? FelV ? ?

Reference Martin, PNSA 1981 Steele, Science 1984 Bonner, PNAS 1982 O’Connell, Virology 1984 Callahan, PNAS 1982 Ono, J. Virol. 1997 Kroger, J. Virol. 1987 Mager, J. Virol. 1987 Perl, Genomics 1991 Noda, Nucl. Acids. Res. 1982 Werner, Virology 1990 La Mantia, Nucl. Acids Res. 1991 Levy, J. Gen. Virol. 1990 Horwitz, J. Virol. 1992

ERV, endogenous retrovirus; HERV-K, human endogenous retrovirus K; HuRRS-P, human repeated retroviral sequence-P; RTVL-H, retroviral-like sequence-H; HRES-1, human retroviral endogenous sequence 1; MoMuLV, Molony murine leukaemia virus; BaEV, baboon endogenous virus; SSAV, simian sarcoma-associated virus; FelV, feline leukaemia virus.

sequence, thus indicating another evolutionary pathway within eucaryotes [8]. One important feature concerning retrotransposable elements others than retroviruses, which should be outlined for this review, is their absence of infectivity, even if some of them can produce intracellular ‘capsid-like’ particles. The acquisition of an env gene by one of these elements (gypsy) has been shown in Drosophila melanogaster as has the infectivity of the resulting particles, thus illustrating a possible retroviral ontogeny from such elements in a eucaryotic cellular genome [9]. However, in addition to this ‘endocellular’ evolutionary pathway, many examples in animal retrovirus families strongly suggest that a horizontal ‘interspecies’ transmission of infectious particles also occurs [10]. The corresponding retroviral genomes which may have emerged from an endogenous retroviral evolution in the productive animals are nonetheless ‘exogenous’ infectious retroviruses for the ‘xenorecipient’ species. A nonlethal integration of such xenotropic strains into the germ cell lineage of newly infected species could thus explain the appearence of newly acquired ERV familes after the evolutionary divergence of particular animal families. For example, ancestral copies of a human retrovirus family, HERV-K, are present in DNA of superior primates (Old World monkeys and humans), but not in New World monkeys and inferior animal species [11]. According to these concepts, the existence of exogenous retroviral genomes such as HIV-1 and HTLV-1 with very limited homologous ERV sequences in the human DNA raises the question of the origin of their own genes. The simian origin of the ancestral strains corresponding to these human retroviruses probably constitutes an example of a recent interspecies transmission without alreadyreported integration in the human germ cell lineage [12]. 2.4. Coding capacity of endogenous retroviral elements

Whatever their particular origin [13] could be, ERVs have not encountered the genetic selection common for Microbes and Infection 1999, 309-322

cellular genes, since they are not associated with a vital or selective function for the corresponding species (apart from an incidental evolution towards a ‘domestication’ of particular ERV genes [5]). Consequently, most ERVs are transcriptionally inactive following mutations (stop codons), deletions and/or various recombination events. In addition, multiple retrotransposition events have usually generated the presence of more-or-less truncated copies in the cellular genomes, in numbers ranging from a single copy to hundreds or even thousands (cf. table I). Particular ERV proviral copies have conserved a long terminal repeat promoter activity and generally partial open reading frames (ORF), thus explaining the possible detection of more-or-less defective ERV RNA in normal cells [14, 15]. Particular ERV RNAs can be transcribed as more-or-less truncated retroviral proteins. These ERV proteins may therefore be recognized as ‘self-antigens’ by the immune system, if their site and period of expression allows acquired tolerance by immunocompetent cells as for regular autoantigens (cf. figure 2). A limited number of ERV proviruses can produce complete retroviral particles which may be infectious (e.g., ecotropic, amphotropic, or xenotropic strains from MLV ERV proviruses) and therefore may have conserved complete ORFs and replication competence. In humans, such particle production has been observed in placenta and in certain tumour cell lines [16, 17]. In the case of human placenta, the genome associated with such retroviral particles apparently belongs to the HERV-K family [18]. The existence of steroid-response elements in the promoter region of the HERV-K family suggests a role for hormonal modifications during pregnancy in the activation of relatively replication-competent HERV-K copies. These particles are a priori noninfectious, and the corresponding ERV antigens are apparently tolerated by the mother’s immune system as well as foetal antigens. The production of retroviral particles associated with retroviral sequences closely related to ERV families in 311

Review

312

Perron and Seigneurin

Microbes and Infection 1999, 309-322

Figure 2. Replicative retroviruses related to multiple homologous endogenous copies: specific features of virion-producing cells. In a healthy cell, multiple copies related to a family of ERVs are present in the DNA of the nucleus in a more or less fragmented form; a few of these copies can be transcribed in a ‘cellular’ RNA form; a few of these RNA can be translated into protein which will be tolerated as a ‘self’ antigen (autoantigen), if its expression is physiological and allows thymic lymphocyte selection. In a cell harbouring an activated replicative strain homologous to an ERV family (e.g., exogenous MMTV or endogenous polytropic MLV), with a reverse transcription cycle producing extracellular virions, the specific provirus is generally expressed in parallel with homologous ERVs. However, the RNAs of the replicative strain allow the synthesis of all the proteins necessary to the formation of virions and the encapsidation of specific ‘viral’ RNA, but also of closely homologous or even unrelated ERV RNA; they also allow the synthesis of functional enzymes (e.g., reverse transcriptase, protease, integrase). The virions thus released extracellularly are detected by reverse transcriptase activity tests, RT-PCR, even electron microscopy (e.g., negative staining after concentrating particles, or on cross-sections at the surface of productive cells). Nonstructural antigens can also be produced in the cell; some of them can be immunogenic, can have the properties of a superantigen, or can be a cytotoxic protein. An immune tolerance to certain ERV-related proteins can elsewhere inhibit the immune response towards corresponding regions of ‘infectious’ retroviral antigens (total or partial absence of response); conversely, the chronic presentation of ‘infectious’ antigens in an inflammatory context can induce a breakdown in the tolerance of homologous ‘autoantigens’ from ERVs expressed by healthy cells (autoimmunity). The possible presence of a retrovirus superantigen can also set off and/or potentialize the immunopathological phenomena.

Retroviral particles and sequences in autoimmune diseases

human tumour or leukaemia cells [19] could result from abnormal (re)activation of virion-coding ERV copies with intact ORFs, in association with the tumoral process. However, the possibility that these retroviral particles could result from an oncogenic process driven by a pathogenic strain, eventually recombined with related ERVs or becoming defective in the posttransformation period, should not be excluded. As for placental particles, no infectivity has been demonstrated for such tumourassociated retroviral particles. This absence of infectiousness can usually be explained by mutations or deletions within the genes involved in virion maturation, reverse transcription, or provirus integration, or simply by the absence of env-encoded motifs permitting a cellular receptor-mediated entry [20]. Nonetheless, the noninfectious nature of these retroviral particles does not exclude possible genetic reversion towards an infectious retrovirus from more-or-less replicationcompetent but nonpathogenic ERVs.

3. Retroviral particles and sequences in autoimmune diseases Diseases are usually classified as ‘autoimmune’ when elevated autoantibody titres and/or chronically activated autoreactive T lymphocytes are encountered in patients and no particular aetiological agent has been identified. Beyond the serological results and animal models which have provided arguments in favour of a potential role for retroviral agents in autoimmune diseases [4], data concerning the detection of extracellular retrovirus-like particles, their association with reverse transcriptase activity, and their in vitro transmissiblility have been reported in different human diseases. In these autoimmune diseases, virions or retrovirus-like particles seem to be specifically produced by certain cells from infected patients, but not by cells of the same phenotype from healthy controls (see below). The production of these ‘retroviral particles’ in nontumoral cells cannot a priori be attributed to a physiological and/or ubiquitous phenomenon, since various stimulations (immunological, chemical, or viral) of phenotypically identical cells from healthy controls did not induce such particle production. Such retrovirus-like particles have been reported in several human autoimmune diseases and are now tentatively associated with particular retroviral genomes. 3.1. Rhumatoid arthritis

Extracellular particles compatible with a retroviral morphology have been observed in synovial fluids from patients with rhumatoid arthritis [21]. In an independant study, a reverse transcriptase activity has been detected in culture supernatants of synoviocytes taken from affected joints in patients with rhumatoid arthritis [22]. A recent study has tried to characterize a retroviral genome associated with rhumatoid arthritis [23], but the use of degenerate primers on samples containing cellular DNA and RNA yielded a wide range of numerous ERVs by polymerase chain reaction (PCR) amplification. Possible candidate sequences overrepresented in rhumatoid arthritis samples Microbes and Infection 1999, 309-322

Review

versus controls were dicussed by theses authors but, in the absence of an appropriate preparation of the RNA encapsidated in these particles, free of any trace of cellular DNA and RNA, no clear association was possible. Indeed, this study did not address the question of retroviral genome identification in rhumatoid arthritis by obtaining a virion-producing cell culture and subsequent selection of extracellular virion-associated RNA, as in the following examples. 3.2. Sjögren’s syndrome (SS) and systemic lupus erythematosus (SLE)

Retroviral particles associated with reverse transcriptase activity and antigens recognized by anti-HIV monoclonal antibody (against capsid protein) have been isolated by Garry et al. from SS patient samples [24]. These virions have thus been named HIAP (human intracisternal A-type particles) on the basis of the morphology of intracellular particles observed by electron microscopy. The same authors further showed a serological recognition of these antigens by the antibodies of patients with SS, SLE, Still’s disease, and idiopathic T-lymphocytopenia (CD4+) [25, 26]. However, the experimental results showing in vitro transmissibility [24], would rather suggest an affiliation to a subfamily other than that of the IAPs (retrotransposons, theoretically nontransmissible). Retrovirus isolation in similar conditions has been reproduced by Griffiths et al., who succeeded in characterizing the pol region of a retroviral genome, provisionnally called HRV-5, from virion-associated RNA [1]. Analysing this pol sequence has potentially revealed a new human retrovirus related to type B or type D oncoviruses. As the sequence of the entire genome has not been cloned, the phylogenic classification is difficult. The authors have argued they were of exogenous origin. Given the data summarized in the first part of this review, if HRV-5 is an oncovirus related to type B or D, it is quite surprising that no ERV homologues were found in the human genome, all the more with the relatively conserved pol gene – unless HRV-5 corresponds to a recent interspecies transmission of a xenotropic retrovirus from animal to human, which would thus have little or no ERV homologues in the human DNA. Whatever its origin will finally reveal, the technological difficulties linked to the very small quantity of HRV-5 retrovirus present in the biological samples have so far made it impossible to obtain its complete sequence. The fact that this sequence appeared to be exogenous in humans has enabled the authors to carry out PCR studies directly on DNA, which is not possible when the homologous ERVs exist in host cells (for example, multiple sclerosis-associated retrovirus [MSRV] or IDDMK, see below). By PCR, the HRV-5 pol sequence was detected in the culture of salivary gland tissue in one case of SS and in three patients without SS operated for head tumours, in spleen tissue from another case of SS, and from lymph glands in a case of SLE. The sequencing of the amplified products confirmed the specificity of the technique but also eliminated a phenomenon of random contamination in view of interisolate sequence variations [1]. In the absence of healthy controls and of large series of patients 313

Review

with untreated SLE or SS, the association of the HRV-5 retrovirus with these diseases remains unclear. In addition, more recent results obtained with HRV-5 are apparently not consistent with a clear association with SS or SLE [27], though technical limits due to an infection of a very restricted number of cells in vivo and/or to genetic variations cannot be ruled out for the moment. The improved detection of the HRV-5 sequence reported by these authors after cell culture and virion concentration on sucrose gradient [1] may indicate the necessity to culture the patients’ tissues prior to PCR analysis. Neither are these preliminary PCR results consistent with those of the serological studies performed by Garry et al. with HIAP isolates [24-26]. Therefore, the identity between the HIAP virion and the HRV-5 genome remains to be confirmed. The epidemiology of HRV-5 itself, which appears to be a novel human exogenous retrovirus, remains to be studied in large series of patients and healthy individuals. Nonetheless the possibility that an exogenous retrovirus, different from HIAP and not associated with SS or SLE, could ‘superinfect’ some patients with these diseases cannot be excluded, as heterologous retroviruses have been found in association in the same disease [28]. Alternatively, infectious retroviruses can drive the propagation of endogenous and even defective retroviral RNAs, which nonetheless carry the determinants of pathogenicity [29]. The identification of an HIAP-associated genome would still have to be performed if HRV-5 detection was thus made in the context of a dual retroviral expression (e.g., helper/defective, as in murine AIDS) or was only a coincidence in SS and SLE (e.g., the codetection of visna retrovirus in ovine pulmonary carcinoma caused by JSRV). Finally, even if it was not proven to be particularly associated with SS or SLE, HRV-5 infection in humans should be studied. For HRV-5 presence in particular individuals is presently unexplained. 3.3. Type 1 diabetes, or insulin-dependent diabetes mellitus (IDDM)

Conrad et al. have shown the existence of a reverse transcriptase activity in primary culture supernatants of pancreatic islets biopsied from patients with IDDM [2]. Beforehand, they had shown a superantigen effect of these cultures on human lymphocytes [30]. In the supernatants of β-islet and leukocyte cultures from diabetic patients, the same authors have succeeded in cloning overlapping regions of a retroviral genome associated with extracellular particles, and named it IDDMK [2]. This retrovirus belongs to the group defined by the human ERV prototype HERV-K (table I). The endogenous origin of this IDDMK retrovirus strain seems established according to these authors. By RT-PCR, the IDDMK was detected in the plasma of 10 patients with IDDM in primary decompensation [2]. The absence of detection in the plasma of 10 healthy control subjects suggested an association with IDDM, which has to be confirmed on larger series. The expression of this retrovirus and the consecutive production of extracellular virions is apparently not linked to the simple stimulation of lymphoid or pancreatic cells by the inflammatory process: indeed, after mitogenic 314

Perron and Seigneurin

stimulation, induction of cell death, or coculture with allogenic lymphocytes, neither virion production nor reverse transcriptase activity was observed [2]. According to these authors’ results, retrovirus expression takes place in leukocytes, and therefore, the known transcription of HERV-K copies in activated lymphocytes would not account for the production of extracellular particles associated with superantigen activity, in healthy controls at least [2, 30]. The potential pathogenicity of these virions was supported by the study of env gene sequences: an ORF coding for a superantigen was identified. Its expression in cells transfected by a plasmid containing IDDMK superantigen ORF reproduced the clonal stimulation of Vb7+ human T lymphocytes, characteristic of the superantigen effect already observed [2]. The existence of a superantigen encoded by an ERV had been demonstrated for the first time in the case of the murine retrovirus MMTV [31]. The ORF for the MMTV superantigen is located in the U3 region, normally noncoding, of the 3’ long terminal repeat. However, even though the HERV-K family is related to MMTV, the reported IDDMK superantigen is not encoded by this U3 region but by the env gene ; neither is it significantly homologous in amino acid sequence with MMTV superantigen [2]. Recent studies have failed to confirm a significant detection of IDDMK RNA in type 1 diabetes by a nonnested RT-PCR [32, 33]. In addition, another group could not detect superantigen activity with another HERV-K clone differing from the described IDDMK by six amino acids [34], since the original clone was not available. These groups also discussed the exact origin of the IDDMK RNA amplified in β-islet cultures from diabetic patients. These results raise questions about i) the difficulty in standardizing and monitoring PCR techniques when primers can codetect numerous endogenous copies from the same genetic family, ii) the genetic polymorphism and diversity of particular elements within an ERV family – but also about their possible de novo retrotransposition, iii) the possible complementation and/or recombination of RNAs from different proviruses, and iv) a possible role of limited amino acid variations in the superantigen activity described for IDDMK superantigen. Although further studies are necessary, the role of a retroviral superantigen in the genesis of an autoimmune process leading to the inflammatory destruction of pancreatic β-islets in type 1 diabetes is still plausible, according to Conrad [35]. Mechanisms through which a superantigen can induce the activation and the uncontrolled proliferation of T- and B-lymphocyte clones against various autoantigens have been described [36] and are represented schematically in figure 3. Though molecular mimicry hypotheses have long been considered among the best explanations for the induction of autoimmunity by microbial agents, the role of superantigens in autoimmunity now appears critical, in particular when such (exogenous or endogenous) retroviruses may prove to be associated with an autoimmune disease. Microbes and Infection 1999, 309-322

Retroviral particles and sequences in autoimmune diseases

Microbes and Infection 1999, 309-322

Review

315

Figure 3. Schematic representation of lymphocyte activation by a superantigen. A ‘normal’ antigen can activate T lymphocytes only i) if associated with a major histocompatibility class II molecule at the surface of an antigen-presenting cell (e.g., macrophage) and ii) if specifically recognized by the antigen recognition site on the T-cell receptor molecule. In these ‘classical’ conditions, an antigen will not activate more than 0.01% of T lymphocytes. In the case of a superantigen, the T-cell receptor does not ‘recognize’ a specific antigen, but is itself ‘specifically recognized’ by this superantigen, due to superantigen affinity for particular motifs on certain T-cell receptor variable chain (Vβ) subtypes. A given superantigen is specific for a particular Vβ ‘x’ type expressed in T-cell receptors from a variable percentage of T lymphocytes, but this recognition is independent of the antigenic specificity recognized by the corresponding T-cell clones. In addition, superantigens can bind class II major histocompatibility complex molecules (DR antigens) expressed on antigen-presenting cells at distance from the ‘peptide groove’, on the external part of the molecule, as for the T-cell receptors. The cross-linking thus performed by a superantigen between these surface molecules of the two cell types (antigen-presenting cells and T cells) normally co-operating the specific immune response, induces activation and, in first challenges, clonal expansion of any Vβ ‘x’ T cells. The proportion of stimulated cells in the presence of a superantigen can consequently be as high as 10 to 40% of T lymphocytes and provoke massive cytokine secretion totally disproportionate with the pathogen load (bacteria, virus, endogenous retrovirus...). A systemic production (e.g., Staphylococcus enterotoxin B) will cause ‘septic shock’ essentially due to tumour necrosis factor-alpha overproduction in plasma. A tissue-localized superantigen expression (e.g., ERV in a few isolated cells) would cause an acute and very excessive inflammation in the presence of very few productive cells. The phases of hyperstimulation by superantigens are generally followed by phases of specific anergy of T-lymphocyte clones, if repeatedly stimulated by a given superantigen, and can end by apoptosis-induced specific deletion of a T-cell repertoire bearing the recognized Vβ ‘x’ domain.

Review

3.4. Multiple sclerosis (MS)

In multiple sclerosis, retrovirus-like extracellular particles associated with a reverse transcriptase activity were detected by different groups. We first described such particles in a primary culture of leptomeningeal cells (LM7) from the cerebrospinal fluid of a patient with MS [37]. The particles and the reverse transcriptase activity produced in the culture supernatant were not found in the control cells. In addition, the polygenic stimulation by phorbol esters and butyric acid or the superinfection by the herpes simplex virus type 1 increased retrovirus production in the supernatant of the LM7 culture, but induced neither particle production nor specific reverse transcriptase activity in the control cultures [37]. Later, this activity and these retrovirus-like particles were found in series of monocyte cultures from patients with active MS but not in the supernatants of cultures from healthy controls, from neurological diseases other than MS, and from MS in remission for more than a year [38]. S. Haahr’s team in Denmark made similar observations in cultures of B lymphocytes spontaneously immortalized in vitro by Epstein-Barr virus [39]. More recently, another team has established lymphoid cell lines from MS patients, producing retrovirus-like particles associated with a reverse transcriptase activity [40]. We had considered, by that time, that this retrovirus could be endogenous [41] but that, according to our results, the provirus producing extracellular particles may not be present or may not be replication-competent in the genome of the non-MS controls already tested. Our strategy for the molecular characterization of this retroviral expression was primarily based on the RNA genome encapsidated in these particles. In fact, only particle-associated RNA could permit a preliminary identification, since normal cellular RNA and DNA could contain homologous but not relevant ERV genomes. Using material from previous cultures, we were able to produce and purify extracellular particles for a first study by RT-PCR combined with extensive DNase treatments of samples, buffers, and enzyme cocktails [42]. The primers used corresponded to two conserved regions in the retrovirus pol gene encompassing a variable region of approximately 100 base pairs [42]. The results of this study have enabled, apart from irrelevant primer concatemers or fragments shorter than the expected size, the identification of sequences corresponding to the expected pol region [3]: a preponderant group (MSRVcpol, representing 85–90% of adequate clones) with about 70% of homology with a human ERV (ERV9) [43], and a minor group (10–15% of adequate clones) with sequences really homologous (over 90%) to ERV9. The same technique later confirmed that these sequences were present in the supernatants of productive MS cell cultures and coincided with the sucrose density gradient containing particles associated with a reverse transcriptase activity [3]. With RT-PCR extension approaches on concentrated extracellular particles, we have obtained the sequence of the complete protease and RT sequences [3] contiguous to the first MSRVcpol fragment. The virus producing particles containing such retroviral RNA was provisionally called MSRV owing to the 316

Perron and Seigneurin

origin of its isolation. By RT-PCR extensions, we have been able to progress beyond the 5’ and 3’ ends of this MSRV ‘pol’ fragment on the RNA encapsidated in such particles and have currently obtained clones overlapping on the expected regions of a retrovirus genome [44]. As illustrated in figure 4, the phylogenic analysis of the pol gene enables this gene to be linked to the type C oncovirus genes. The phylogenic analysis of the MSRV env gene, however, indicates closer proximity with the type D retroviruses (not shown). This phylogenic divergence between retroviral pol and env genes producing virions with type D morphology has already been described and interpreted as deriving from an ancestral recombination [45]. These phylogenic observations are compatible with the morphology of particles observed in the LM7 cultures [37] and in monocyte cultures [38]. At present we have not been able to clone a replicationcompetent MSRV provirus which could encode such particles and cannot easily achieve this objective, since many ERV copies with strong nucleotide homology with every MSRV region exist in human DNA. We have indeed identified a novel human ERV family (HERV-W) [46], different from ERV9 and genetically related to MSRV sequences, as confirmed by phylogenic analyses and particular retroviral features. For example, the primer binding site identified in MSRV RNA 5’ sequences [44, 46] corresponds to the tryptophane (W) t-RNA, while in the ERV9 family, it corresponds to an arginine (R) t-RNA [47]. However, the significance of a minority of apparently copackaged ERV9 sequences in our virion isolates [3] is unclear. Their presence in low numbers relative to MSRV sequences may not a priori rule out a possible biological role of such particle-associated defective ERV9 RNAs, as recently objected [48], if interfering with retroviral expression and/or if expressed as partial proteins with biological effects. For example, such cases are encountered in murine AIDS, caused by the association in retroviral particles, of a defective MLV gag RNA encoding pathogenic determinants and of a helper retrovirus genome, encoding infectious retroviral particles [29]. A previous study had shown that the conserved ‘panretro’ pol fragment corresponding to ERV9 and ERV9-like RNA could be amplified in cellular RNA extracted from MS brain tissues [49]. However, Northern blot analysis performed with this small and conserved fragment in rather low-stringency conditions showed ubiquitous and possibly nonspecific hybridization in human multitissue RNA preparations [49]. In the cytoplasm of cells obtained from non-MS controls, we have also detected RNA homologous to MSRV sequences (HERV-W) expressed in small quantities, using RT-PCR with nondegenerate primers. However, a study on cerebrospinal fluid cell RNA with the pan-retro RT-PCR technique followed by specific MSRV ELOSA hybridization showed, under these conditions, no detectable MSRV product in several controls with numerous activated lymphocytes in cerebrospinal fluid (e.g., viral encephalitis), whereas a positive signal was found in five out of 10 samples corresponding to untreated MS patients [3]. As previously evoked, RT-PCR with MSRV primers is, however, not discriminant within cellular or tissue RNA, since baseline expression of MSRV-related Microbes and Infection 1999, 309-322

Retroviral particles and sequences in autoimmune diseases

Review

Figure 4. Phylogenic tree (pol gene). The tree was calculated with GeneWorkst software (UPGMA tree) based on the amino acid sequence encoded by the most conserved region of retroviral pol gene (VLPQG....YXDD region). HSRV, human spumaretrovirus; EIAV, equine infectious aenemia virus; BLV, bovine leukaemia virus; HIV1 and HIV2, human immunodeficiency viruses type 1 and 2; HTLV1 and HTLV2, human leukaemia viruses type 1 and 2; F-MuLV, Friend-murine leukaemia virus; MoMLV, Moloney-murine leukaemia virus; BAEV, baboon endogenous virus; GaLV, gibbon ape leukaemia virus; HUMER41, human endogenous retroviral sequence clone 41; IAP, intracisternal A-type particle; MPMV, Mason-Pfizer monkey virus; HERVK10, human endogenous retrovirus ‘K10’; MMTV, mouse mammary tumour virus; HSERV9 (reference ERV9 clone from cellular RNA), human sequence of endogenous retrovirus 9; MSRV, multiple sclerosis associated retrovirus; SIV, simian immunodeficiency virus; RTVL-H, retrovirus-like sequence ‘H’; SFV, simian foamy virus; VISNA, visna retrovirus; SIV1, simian immunodeficiency virus type 1; SRV-2, simian retrovirus type 2; SMRV-H, squirrel monkey retrovirus ‘H’. Microbes and Infection 1999, 309-322

317

Review

(HERV-W) transcripts can be amplified with MSRV primers tested today in different genes [45] (cf. general principles described in figure 2). Its sensitivity is much greater than the ‘pan-retro’ RT-PCR, in which degenerate consensus primers create competitive amplification of any intracellular ERV RNA and which only allows detection of major retroviral RNA population(s) in a tissue sample. Nevertheless, using an RT-PCR technique which selects the MSRV pol sequence packaged into extracellular particles, and not cellular or tissue RNA, we have recently obtained results suggesting that circulating ‘particle-associated’ MSRV RNA was detectable in MS sera, preferentially during untreated and active periods [50]. These results therefore suggest that, similarly to MS cell cultures, an in vivo production of extracellular retroviral particles may occur in patients during particular periods. However, optimization of the present complex technical procedures, addressing very low particle numbers and requiring careful hydrolysis of nonpackaged RNA and DNA without destruction of particle-associated RNA, is required for further studies and confirmation. Indeed, our recent evaluations suggest that, in the present conditions, only statistical detection of low numbers of circulating particles can be obtained. This actually precludes perfect reproducibility and standardization of such a nested RT-PCR protocol, in the absence of major technical improvements. From our present data, MSRV RNA may reflect the expression of i) a human endogenous retrovirus (HERV-W) haplotype particular to MS DNA, competent for replication, encoding viral particles and possibly triggered by a heterologous viral transactivator (e.g., herpesvirus), ii) an exogenous proviral DNA, member of a genetically homologous HERV-W family and interfering with endogenous copies at the origin of defective RNA coencapsidated in extracellular particles, iii) a stimulation and rather selective coencapsidation of particular HERV copies with a pol gene phylogenically related to the HERV-W/ERV9 cluster of type C oncoviruses, by a ‘helper’ retrovirus, or iv) multicomplementation of partially defective HERV copies transactivated by (viral) cofactors. Indeed, the very recent publication by S. Haahr’s group in Denmark, of sequence data from four MS B-cell line isolates [51] also suggests copackaging of RNAs from a cluster of closely related ERVs having pol sequences related to type C oncoviruses. The gag and env ‘RGH’ sequences they report in particle-associated RNA belong to the retroviral-like sequence H (RTVL-H) family [52]. As shown in figure 4, MSRV, ERV9 (HSERV9 clone), and RTVL-H pol sequences define a common phylogenic ‘cluster’. According to most data obtained in similar cases of retroviral copackaging [53], the retroviral genome encoding theses particles has the greatest probability of belonging to the major family of encapsidated retroviral RNAs; unless a heterologous ‘helper’ retrovirus is involved in particle production with full-length coding RNA encapsidated in a minority of ‘wild-type’ particles only [53]. As previously described, our nonselective ‘pan-retro’ RT-PCR approach has amplified retroviral sequence subgroups in similar proportions (85–90% of ‘MSRV-cpol’ and 10–15% of ERV-9) from different cultured cell-types [3]. Since 318

Perron and Seigneurin

RGH/RTVL-H pol sequences were not encountered among about dozens of clones sequenced from our three different culture isolates, they may have been present in too low numbers to have a chance of being readily sequenced in our pan-retro series. If so, this may reflect a lower affinity for the retroviral particles produced in MS and/or a lower stimulation in virion-productive cells. Alternatively, the pan-retro primers may not have efficiently amplified this particular HERV family (RTVL-H). Since such a ‘nonselective’ approach has not been used by this group, their specific primers could easily amplify a small subpopulation of copackaged RNAs. However, fluctuations in the proportions of such copackaged ERV RNAs may occur during passages in cell cultures and in vivo, during disease evolution. Therefore, quoting T. Christensen et al. [51], ‘the fact that RGH and MSRV belong to the same subsection of the type C Oncovirinae may indicate that we are looking at two features of the same phenomenon’. Contrary to similar observations obtained with a panretro RT-PCR approach which showed copackaging of two unrelated ERV families of sequences in retroviral particles from a human breast cancer cell-line [54], the retroviral sequences detected by different aproaches in ‘MS particles’ belong to closely related elements (MSRV, HSERV9 and eventually, RTVL-H; cf. figure 4). This may reflect a common transactivating factor and/or a common – and relative – affinity for RNA packaging signals encoded by a particle-encoding genome from this phylogenic cluster [53]. Studying the potential role in MS pathogenesis of such retroviral expression, therefore, requires further confirmation of the retroviral genome(s) contributing to retroviral particle and RT-activity production specific to MS cells. The search for a superantigen associated with this retroviral production is also of importance, since it had already been invoked as a possible explanation for the immunopathological process observed in MS [55]. In parallel, the study of a gliotoxic protein coexpressed with retroviral activity in MS macrophage cultures [56] and further detected in patients’ cerebrospinal fluid and urine [57, 58] may also provide a direct or indirect link with the disease [59]. Finally, even if retroviral expression could be a consequence of such a disease rather than a primary cause, associated retroviral sequences, biological activities, or antigens could nonetheless constitute useful biological markers of the disease. But, if they only represented a nonspecific consequence of immune dysregulation, it should be understood anyway why retroviral particles observed in other autoimmune diseases such as ES, SLS, or IDDM are apparently associated with other retroviral genomes or HERV families than those found in MS.

4. Conclusion This review and the numerous but partial data reported here show that new questions, possibly pertinent but very complex, have been addressed with recently developed techniques. They have obviously not been answered Microbes and Infection 1999, 309-322

Retroviral particles and sequences in autoimmune diseases

clearly and definitely. Nonetheless, the partial but converging data already gathered now provide tools which can be used to further explore such important questions. A common point in these approaches is the identification of retroviral genomes at the level of retrovirus-like particles produced by cell cultures from patients with several autoimmune disease. Because of very low and/or transient production, these approaches have used RT-PCR amplification on ‘virion-associated’ RNA. This molecular approach was nonetheless preceeded by steps of concentration and purification of extracellular particles associated with RT activity in culture fluids. This approach, rather than a strategy based on intracellular RNA differential expression, appears to be the most appropriate one since it avoids codetection of numerous and irrelevant ERV sequences from intracellular DNA or RNA. Indeed, ERVs expressed at sufficient levels in nonproductive cells in a tissue or a cell culture can mask particular RNA expression associated with ‘virion-production’ in few cells only. This is all the more necessary when the first possible approach is based on consensus primers or when homologous ERV RNAs exist in host cells and can be amplified with nondegenerate primers. The necessity of eliminating all possible trace of contaminating DNA and of nonencapsidated RNA in particle-associated RNA extract, is also a prerequisite for targeting retroviral genomes which have ERV counterparts in the host cell genome. But, even in the best situations, these precautions cannot preclude detection of copackaged RNA in retroviral particles. This phenomenon has been well reviewed by Linial et al. [53] and, in case particles are encoded by an exogenous retrovirus with an ERV counterpart, homologous ERV RNA copackaging can compete quite strongly with the autologous ‘wild-type’ RNA genome. As part of the complex ‘biological life’ of such retroviruses, it also appears necessary to study copackaged ERV genomes which may account for their potential pathogenicity by e.g., recombinations [60] or propagation of defective clone expressing pathogenic molecules [29], and may be at the origin of their rapid loss of infectivity by defective interference and/or ERV takeover [53]. The complexity of retroviral genome studies in these situations, represented in this review by IDDMK in autoimmune diabetes and MSRV in multiple sclerosis, can become a major difficulty for a definite conclusion. Finally, obtaining specific antibody against recombinant proteins encoded by these genomes is also required in order to verify that such retrovirus-like particles can be immunolabeled, therefore confirming the genetic origin of their proteins, in case different retroviral RNAs can be copackaged. In each autoimmune disease discussed here (rhumatoid arthritis, SLE, SS, IDDM, or MS), the identification of the correct proviral clone at the origin of these particles, the search for cytotoxic molecules and/or superantigens associated with a specific retroviral expression in these diseases, as well as the study of immune tolerance induction or breakdown against autoantigens, including eventual HERV proteins, are therefore converging approaches. It may take a long time before the knowledge on such human retroviruses, though analogous to well-studied animal oncoviruses, can help elucidating a pathogenic role of Microbes and Infection 1999, 0-322

Review

exogenous or endogenous human oncovirus genomes. However, progress in the general knowledge of these retroviral entities in humans is certainly needed. In parallel, other studies which were not primarily focused on ‘extracellular’ retroviral particles in cell cultures, have also suggested a possible association of a spumavirus genome in Graves’ disease [61], which was not found by others [62] but was recently shown to possibly involve particular spumavirus genotypes [63]. Other recent data also demonstrate that HTLV-1 can induce the ‘specific’ production of autoantibodies reacting with neuronal cells in HAM/TSP patients [64]. Therefore, the perspectives opened by the concept of a retroviral aetiopathogenesis involving different retroviruses – possibly encoding potent immunopathogenic molecules such as superantigens and/or causing tolerance breakdown or even molecular mimicry – in autoimmune diseases, are considerable but now require increased and concerted research in order to evaluate their reliability.

Glossary i) Exogenous retrovirus: an exogenous retrovirus is transmitted by ‘classical’ infectious routes. Its proviral copies are present in the DNA of infected cells only and consequently are not present in the DNA of every cell in an infected individual and in none of a noninfected one. ii) Endogenous retrovirus (ERV): an endogenous retrovirus is transmitted by the gametes according to ‘classical’ mendelian genetics. If genetically inherited by an individual, its provirus copies are present in the DNA of every cell of this individual. iii) Tropism of endogenous murine leukaemia viruses (MLV): several endogenous MLV strains produce virions which can infect cells of various species in vitro; they are classified as: a) ecotropic, if they can infect cells from the original species only. b) xenotropic, if they can infect cells from other species only (and not of the original species producing the virions). c) amphotropic, if they can infect cells from the original species as well as that of another species. d) polytropic, if they can infect cells from various species. iv) The classification of ERV families can be based on the type of tRNA complementary sequence present in the primer binding site of their members. For example, the HERV-K family is a human ERV family in which copies carrying an intact primer binding site region have sequences complementary to lysine-binding tRNA sequence. v) Retrotransposons and Retroposons: the key feature which distinguishes them from retroviruses is the lack of an env gene consistent with an intracellular replication cycle and an absence of infectivity. They contain gag and pol genes which are surrounded by LTRs, in the case of retrotransposons, or usually contain a repeated segment at their 5’ end, which is required for their transcription in the case of retroposons (figure 1).

References [1] Griffiths D.J., Venables P.J., Weiss R.A., Boyd M.T., A novel exogenous retrovirus sequence identified in humans, Virology 71 (1997) 2866–2872. [2] Conrad B., Weissmahr R.N., Boni J., Arcari R., Schupbach J., Mach B., A human endogenous retroviral superantigen as candidate autoimmune gene in type I diabetes, Cell 25 (1997) 303–313. 319

Review

[3] Perron H., Garson J.A., Bedin F., Beseme F., ParanhosBaccala G., Komurian-Pradel F., Mallet F., Tuke P.W., Voisset C., Blond J.L., Lalande B., Seigneurin J.M., Mandrand B., The Collaborative Research Group On Multiple Sclerosis, Molecular identification of a novel retrovirus repeatedly isolated from patients with multiple sclerosis, Proc. Natl. Acad. Sci. USA 94 (1997) 7583–7588. [4] Nakagawa K., Harrison L.C., The potential roles of endogenous retroviruses in autoimmunity, Immunol. Rev.152 (1996) 193–236. [5] Renault S., Bonnin-Rouleux F., Periquet G., Eléments transposables et évolution des génomes, Virologie 1 (1997) 133–44. [6] Temin H.M., Origin and general nature of retroviruses, in: Levy J.A. The retroviridae, New York Plenum Press, NY, (Ed.), 1992. [7] Doolittle R.F., Feng D.F., Tracing the origin of retroviruses, Curr. Top. Microbiol. Immunol.176 (1992) 195–211. [8] Eickbush T.H., Telomerase and retrotransposons: which came first?, Science 277 (1997) 911–912. [9] Kim A., Terzian C., Santamaria P., Pelisson A., PurdHomme N., Bucheton A., Retroviruses in invertebrates the gypsy retrotransposon is apparently an infectious retrovirus of Drosophila melanogaster, Proc. Natl. Acad. Sci. USA 91 (1994) 1285–1289. [10] Todaro G.J., Interspecies transmission of mammalian retroviruses,in: Klein G. Viral Oncology, Raven Press, New York, NY, (Ed.), 1980, pp. 291–309. [11] Steinhuber S., Brack M., Hunsmann G., Schwelberger H., Dierich M.P., Vogetseder ., Distribution of human endogenous retrovirus HERV-K genomes in humans and different primates. Hum. Genet. 96 (1995) 188–192. [12] Gojobori T., Moriyama E.N., Ina Y., Ikeo K., Miura T., Tsujimoto H., Hayami M., Yokoyama S., Evolutionary origin of human and simian immunodeficiency viruses, Proc. Natl. Acad. Sci. USA 87 (1990) 4108–4111. [13] Varmus H.E., Cohen J.C., Hughes S.H., Majors J., Bishop J.M., The origin and structure of endogenous retroviral DNA, Ann. NY. Acad. Sci.354 (1980) 379–383. [14] Hummell D.S., Dooley J.S., Khan A.S., Lawton A.R., Coordinate transcriptional control of murine endogenous retrovirus and Ig genes during B cell differentiation, J. Immunol. 151 (1993) 3131–3139. [15] Kato N., Pfeifer-Ohlsson S., Kato M., Larsson E., Rydnert J., Ohlsson R., Cohen M., Tissue-specific expression of human provirus ERV3 mRNA in human placenta two of the three ERV3 mRNAs contain human cellular sequences, J. Virol. 61 (1987) 2182–2191. [16] Nelson J., Leong J.A., Levy J.A., Normal human placentas contain RNA-directed DNA polymerase activity like that in viruses. Proc. Natl. Acad. Sci. USA 75 (1978) 6263–6267. [17] Patience C., Simpson G.R., Coletta A.A., Welch H.M., Weiss R.A., Boyd M.T., Human endogenous retrovirus expression and reverse transcriptase activity in the T47D mammary carcinoma cell line, J. Virol. 70 (1996) 2654–2657. 320

Perron and Seigneurin

[18] Simpson G.R., Patience C., Lower R., Tonjes R.R., Moore H.D., Weiss R.A., Boyd M.T., Endogenous B-type (HERV-K) related sequences are packaged into retroviral particles in the placenta and possess open reading frames for reverse transcriptase, Virology 222 (1996) 451–456. [19] Boller K., Konig H., Sauter M., Mueller-Lantzsch N., Lower R., Lower J., Kurth R., Evidence that HERV-K is the endogenous retrovirus sequence that codes for the human teratocarcinoma-derived retrovirus HTDV, Virology 196 (1993) 349–353. [20] DeGiuli M., Orghen C., Custode P., Lavanga E., Negri A., Radaelli F., Radaelli A., Biological and molecular analysis of LCV, an endogenous retrovirus with defective env gene, Arch. Virol.102 (1988) 99–110. [21] Stransky G., Vernon J., Aicher W.K., Moreland L.W., Gay R.E., Gay S., Virus-like particles in synovial fluids from patients with rheumatoid arthritis, Br. J. Rheumatol.32 (1993) 1044–1048. [22] Aicher W.K., Stransky G., Gay R.E., Gay S., Synovial lining cells derived from patients with rheumatoid arthritis (RA) produce reverse transcriptase, Arthr. Rheum. 34 (1991) S177. [23] Nakagawa K., Brusic V., McColl G., Harrison L.C., Direct evidence for the expression of multiple endogenous retroviruses in the synovial compartment in rheumatoid arthritis, Arthr. Rheum.40 (1997) 627–638. [24] Garry R.F., Fermin C.D., Hart D.J., Alexander S.S., Donehower L.A., Luo-Zhang H., Detection of a human intracisternal A-type retroviral particle antigenically related to HIV, Science 250 (1990) 1127–1129. [25] Garry R.F., Fermin C.D., Kohler P.F., Markert M.L., Luo H., Antibodies against retroviral proteins and nuclear antigens in a subset of idiopathic CD4+ T lymphocytopenia patients, AIDS Res. Hum. Retrovir.12 (1996) 931–940. [26] Dang H., Dauphinee M.J., Talal N., Garry R.F., Seibold J.R., Medsger T.A. Jr, Alexander S., Feghali C.A., Serum antibody to retroviral gag proteins in systemic sclerosis, Arthr. Rheum.34 (1991) 1336–1337. [27] Rigby S.P., Griffiths D.J., Weiss R.A., Venables P.J., Human retrovirus-5 proviral DNA is rarely detected in salivary gland biopsy tissues from patients with Sjogren’s syndrome, Arthr. Rheum.40 (1997) 2016–2021. [28] Sharp J.M., Sheep pulmonary adenomatosis: a contagious tumour and its cause, Cancer Surv. 6 (1987) 73–83. [29] Kubo Y., Kakimi K., Higo K., Kobayashi H., Ono T., Iwama Y., Kuribayashi K., Hiai H., Adachi A., Ishimoto A., Possible origin of murine AIDS (MAIDS) virus: conversion of an endogenous retroviral p12gag sequence to a MAIDS-inducing sequence by frameshift mutations, J. Virol. 70 (1996) 6405–6409. [30] Conrad B., Weidmann E., Trucco G., Rudert W.A., Behboo R., Ricordi C., Rodriquez-Rilo H., Finegold D., Trucco M., Evidence for superantigen involvement in insulindependent diabetes mellitus aetiology, Nature 371 (1994) 351–355. [31] Marrack P., Kushnir E., Kappler J., A maternally inherited superantigen encoded by a mammary tumor virus, .Nature 349 (1991) 524–526. Microbes and Infection 1999, 0-322

Retroviral particles and sequences in autoimmune diseases

[32] Lan M.S., Mason A., Coutant R., Chen Q.Y., Vargas A., Rao J., Gomez R., Chalew S., Garry R., Maclaren N.K., HERVK10s and immune-mediated (type 1) diabetes. Cell 95 (1998) 14–16. [33] Murphy V.J., Harrison L.C., Rudert W.A., Luppi P., Trucco M., Fierabracci A., Biro P.A., Bottazzo G.F., Retroviral superantigens and type 1 diabetes mellitus, Cell 95 (1998) 9–11. [34] Lower R., Tonjes R.R., Boller K., Denner J., Kaiser B., Phelps R.C., Lower J., Kurth R., Badenhoop K., Donner H., Usadel K.H., Miethke T., Lapatschek M., Wagner H., Development of insulin-dependent diabetes mellitus does not depend on specific expression of the human endogenous retrovirus HERV-K, Cell 95 (1998) 11–14. [35] Con et al. In response to Murphy et al., Löwer et al., and Lan et al. Cell 95 (1998) 16. [36] Acha-Orbea H., Macdonald H.R., Superantigens of mouse mammary tumor virus, Annu. Rev. Immunol.13 (1995) 459–486. [37] Perron H., Geny C., Laurent A., Mouriquand C., Pellat J., Perret J., Seigneurin J.M., Leptomeningeal cell line from multiple sclerosis with reverse transcriptase activity and viral particles, Res. Virol. 140 (1989) 551–561. [38] Perron H., Gratacap, L.A., Lande B., Laurent A., Geny C., Mallaret M., Schuller E., Stoebner P., Seigneurin J.M., Isolation of retrovirus from patients with multiple sclerosis, Lancet 337 (1991) 862–863. [39] Haahr S., Sommerlund M., Moller-Larsen A., Nielsen R., Hansen H.J., Just another dubious virus in cells from a patient with multiple sclerosis?, Lancet 337 (1991) 863–864. [40] Lan X.Y., Zeng Y., Zhang D., Hong M.L., Wang D.X., Zhang Y.L., Feng Z.J., Tang M.H., Feng B.Z., Establishment of a human malignant T lymphoma cell line carrying retrovirus-like particles with RT activity, Biomed. Environ. Sci.7 (1994) 1–12. [41] Rasmussen H.B., Perron H., Clausen J., Do endogenous retroviruses and related genetic elements contribute to the development of inflammatory and degenerative nervous system diseases, Acta Neurol. Scand. 88 (1993) 190–198. [42] Tuke P.W., Perron H., Bedin F., Beseme F., Garson J.A., Development of a pan-retrovirus detection system for multiple sclerosis studies, Acta Neurol. Scand. Suppl.169 (1997) 16–21. [43] Lamantia G., Maglione D., Pengue G., Dicristofano A., Simeone A., Lanfrancone L., Lania L., Identification and characterization of novel human endogenous retroviral sequences prefentially expressed in undifferentiated embryonal carcinoma cells, Nucl. Acids. Res.19 (1991) 1513–1520. [44] Paranhos-Baccala G., Komurian-Pradel F., Bedin F., Ounanian-Paraz A., Passas D., Sodoyer M., Ott C., Rajoharison A., Mallet F., Mandrand B., Perron H., Molecular identification and characterization of a novel retrovirus (MSRV) isolated from patients with multiple sclerosis, J. Neuroimmunol. 90 (1998) (suppl.) 75. [45] Sonigo P., Barker C., Hunter E., Wain-Hobson S., Nucleotide sequence of Mason-Pfizer monkey virus an immunosuppressive D-type retrovirus, Cell 45 (1986) 375–385. Microbes and Infection 1999, 0-322

Review

[46] Blond J.L., Beseme F., Duret L., Bouton O., Bedin F., Perron H., Mandrand B., Mallet F., Molecular characterization and placental expression of HERV-W, a new human endogenous retrovirus family, J. Virol. 73 (1999) 1175–1185. [47] Lania L., Di Cristofano A., Strazzullo M., Pengue G., Majello B., LaMantia G., Structural and functional organization of the human endogenous retroviral ERV9 sequences, Virology 191 (1992) 464–468. [48] Brahic M., Bureau J.F., Multiple sclerosis and retroviruses, Ann. Neurol.42 (1997) 984–985. [49] Lefebvre S., Hubert B., Tekaia F., Brahic M., Bureau J.F., Isolation from humain brain of six previously unreported cDNAs related to the reverse transcriptase of human endogenous retroviruses, AIDS Res. Hum. Retrovir.11 (1995) 231–237. [50] Garson J.A., Tuke P.W., Giraud P., Paranhos-Baccala G., Perron H., Detection of virion-associated MSRV-RNA in serum of patients with multiple sclerosis, Lancet 351 (1998) 33. [51] Christensen T., Dissing-Sørensen P., Riemann H., Hansen H.J., Møller-Larsen A., Expression of sequence variants of endogenous retrovirus RGH in particle form in multiple sclerosis, Lancet 352 (1998) 1033. [52] Hirose Y., Takamarsu M., Harrada F., Presence of env genes in members of the RTVL-H family of human endogenous retrovirus-like lements, Virology 192 (1993) 52–61. [53] Linial M.L., Miller A.D., Retroviral RNA packaging: sequence requirements and implications, in: Swanstrom R., Vogt P.K. Retroviruses strategies of replication, Current Topics in Microbiology and Immunology (Eds.), 157,pp. 1990, Springer-Verlag, Berlin, 125–152. [54] Seifarth W., Skladny H., Krieg-Schneider F., Reichert A., Hehlmann R., Leib-Mosch C., Retrovirus-like particles released from the human breast cancer cell line T47-D display type B- and C-related endogenous retroviral sequences, J. Virol.69 (1995) 6408–6416 . [55] Rudge P., Does a retrovirally encoded superantigen cause multiple sclerosis?, J. Neurol. Neurosurg. Psychiatry54 (1991) 853–855. [56] Ménard A., Amouri R., Michel M., Marcel F., Brouillet A., Belliveaud J., Geny C., Deforges F., Malcus-Vocanson C., Dobransky T., Armstrong M., Lyon-Caen O., Mandrand B., Rieger F., Perron H.,, Gliotoxicity, reverse transcriptase activity and retroviral RNA in monocyte/macrophage culture supernatants from patients with multiple sclerosis, FEBS Lett. 413 (1997) 477–485. [57] Malcus-Vocanson C., Giraud P., Broussolle E., Perron H., Mandrand B., Chazot G., A urinary marker for multiple sclerosis, Lancet 351 (1998) 1330. [58] Ménard A., Pierig R., Pelletier J., Bausa P., Belliveau J., Mandrand B., Perron H., Rieger F., Detection of a gliotoxic activity in the cerebrospinal fluid from multiple sclerosis patients, Neurosci. Lett. 245 (1998) 49–52. [59] Ménard A., Amouri R., Dobransky T., CharriautMarlangue C., Pierig R., Benjelloun N., Cifuentes-Diaz C., Gandhour S., Belliveau J., Gascan H., Lyon-Caen O., Perron H., Rieger F., A Gliotoxic factor in multiple sclerosis, J. Neurol. Sci.154 (1998) 209–221. 321

Review

[60] Obata M., Amanuma H., Harada Y., Sagata N., Ikawa Y., env-Related leukemogenic genes (gp55 genes) of two closely related polycythemic strains of Friend spleen focusforming virus possess different recombination points with an endogenous mink cell focus forming virus env gene, Virology 136 (1984) 435–438. [61] Lagaye S., Vexiau P., Morozov V., Guenebaut-Claudet V., Tobaly-Tapiero J., Canivet M., Cathelineau G., Peries J., Emanoil-Ravier R., Human spumaretrovirus-related sequences in the DNA of leukocytes from patients with Graves disease, Proc. Natl. Acad. Sci. USA 89 (1992) 10070–10074.

322

Perron and Seigneurin

[62] Heneine W., Musey V.C., Sinha S.D., Landay A., Northrup G., Khabbaz R., Kaplan J.E., Absence of evidence for human spumaretrovirus sequences in patients with Graves’ disease, Acquir. Immune Defic. Syndr. Hum. Retrovirol.9 (1995) 99–101. [63] Lee H., Kim S., Kang M., Kim W., Cho B., Prevalence of human foamy virus-related sequences in the Korean population, Biomed. Sci.5 (1998) 267–273. [64] Levin M.C., Krichavsky M., Berk J., Foley S., Rosenfeld M., Dalmau J., Chang G., Posner J.B., Jacobson S., Neuronal molecular mimicry in immune-mediated neurologic disease, Ann. Neurol.44 (1998) 87–98.

Microbes and Infection 1999, 0-322